Auteur :
Bradley
Phil
Année de Publication :
0
Type : Article
Thème : Logiciel non documentaire
Introduction
‘Human-powered search engines’ is perhaps a slightly unfortunate term, since it makes me think of lots of people running around on treadmills providing the energy to keep the servers powered up! However, it’s the term in general use, so we’ll go with it. Essentially it means a search engine which can, and will have its results (or at least the position of its results) affected by human intervention, usually by people rating individual results further up or further down the rankings.
The major search engines that you’ll be familiar with all use complex computer algorithms to work out why one result, or Web page, should rank higher than another. These algorithms will take into account the position of words on a page, the repetition, the number of links pointing to a page and so on. Search engines will make particular claims for their own methodology, all claiming that they provide better and more relevant rankings than the opposition. As a result, each engine keeps its exact ‘recipe’ secret, which is one of the reasons why you receive different results for the same search with each and every engine you use – apart of course from the obvious point that they’re each searching different databases